Prediction in Deformation Analysis of Deep Foundation Pit Group Based on Convolutional Neural Network

Effective monitoring and prediction of data in deep excavation construction are vital for ensuring secure construction practices. To enhance safety protocols and mitigate risks during the deep excavation process, this paper focuses on the Hefei Metro Line 7 Huifu Road Station project. Utilizing collected data, we propose a comprehensive deep excavation deformation prediction model for subway stations. Employing Python programming language, the model establishes a sophisticated network to analyze historical pile top settlements, learn intricate dependencies among multiple monitoring data features, and discern temporal trends for projecting future deformation values. Comparative analyses encompassing Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM), Artificial Neural Network (ANN), and Gated Recurrent Unit (GRU) models are conducted. Experimental outcomes underscore the CNN model's ability to yield higher prediction accuracy for detection data, offering a feasible solution with lower training complexity. This study holds significant implications for advancing construction safety protocols and fostering data-driven research.